ETHORITY®
Task
A defined activity is faster, better or less effort-intensive.
Experiment, field study or workflow observation.Does not establish organisation-wide productivity or profit.
EVIDENCE SYNTHESIS · 14 OCTOBER 2026
Saved hours are not AI value.
Original methods, structured evidence synthesis and proposed application studies. Each has its own question, source edition, method and limits.
STEN & ETHORITY Editorial Team · Method owner Sten Franke
14 Oct 2026 · 15 Oct 2026 · Evidence synthesis of existing research
14 Oct 2026 · 15 Oct 2026 · Evidence synthesis of existing research
ETHORITY
From Saved Hours to Attributable AI Value
What the evidence says about AI productivity, business value and the decisions leaders can defend.
Evidence synthesis · 14 October 2026
FEATURED EVIDENCE STUDY
A faster task is an observation. A business outcome needs an explanation.
The study separates bounded task improvement, reported employee benefit, workflow change and enterprise value. Its original contribution is the maintained record that connects a capacity destination, an outcome acceptor, claim-level evidence and an explicit management decision.
Synthesis without pooling incompatible numbers.
The Saved Hours study is a structured evidence synthesis, not a meta-analysis. It reviews primary research, official surveys, identified consultancy studies and established management approaches selected for their relationship to task performance, worker experience, enterprise outcomes or economic and strategic value.
Different samples, questions, units, countries and periods prevent a single pooled percentage. Survey responses remain attributed observations. Commercial interests in vendor and consultancy studies stay visible. The service calculation is a scenario; governing a method does not establish its empirical performance.
The method builds on Theory of Constraints, benefits-realisation management, productivity J-curve research and work redesign. ETHORITY does not claim to have discovered the general gap between AI investment and value.
Five evidence levels. Five different boundaries.
Worker
People report assistance, time saving, learning or reduced burden.
Employee survey, diary or interview.Does not establish accepted output or audited value.
Workflow
End-to-end performance, quality, capacity or cycle time changes.
Process data, matched comparison or operational study.Does not establish causal financial return.
Enterprise
Respondents or accounts indicate cost, revenue, risk or EBIT impact.
Survey, finance data or portfolio analysis.Does not establish general causality across organisations.
Economic / strategic
Technology and complementary investment change productivity, surplus or market structure.
Economic analysis, longitudinal data or strategy research.Does not establish the outcome of a specific intervention.
Original evidence table · English source figure
Four dimensions that must remain separate.
Conversion stage
S0 Hypothesised → S1 Released → S2 Usable → S3 Destination bound → S4 Enabled → S5 Converted → S6 Attributed
Benefit class
Cash released; cost avoided; throughput and growth; quality and customer; risk and resilience; workforce and learning; strategic option. State the recipient.
Evidence for each claim
E1 Claimed · E2 Measured · E3 Checked · E4 Corroborated · E5 Counterfactual-tested
An accepted outcome at S5 does not upgrade its evidence code.
Management decision
PILOT · REDESIGN · EXPAND · STOP
The action follows the evidence and authority. HOLD is a temporary workflow state. These are explanatory method dimensions, not an automated assessment.
What would make us revise the method?
Prospective applications must test whether reviewers assign stages consistently, usable capacity is actually redeployed, named destinations improve realisation, outcome acceptors prevent false claims and the record changes decisions. Include stopped and negative cases, high-volume operations, knowledge work and workforce or risk benefits.
- Define baseline, permissions, quality, destination, owner and decision rule before a test.
- Compare matched periods or units; preserve case mix and concurrent changes.
- Record measurement burden, reviewer disagreement and unresolved alternative causes.
- Revise fields that consume more effort than their decision value; never require a positive case or testimonial.
Limits that stay visible
Small reversible decisions need proportionate depth. Attribution may remain below E5. Slack can create non-cash value. A record can become bureaucracy. M1 Light, M2 Standard and M3 Full scale with materiality, reversibility and consequence—not the price of a service. No proprietary benchmark result is claimed in this edition.
Inspect the complete worksheet
ETHORITY · 15 Oct 2026
Capacity Conversion Record
Did faster work become an accepted business result? Record the evidence and the responsible decision.
Identity
- Organisation and workflow
- Record owner and outcome acceptor
- AI system, model or agent, version
- Materiality: Light / Standard / Full
- Decision date and next review
Business problem and baseline
- Outcome, recipient, unit and deadline
- Baseline: volume, effort, time, quality, cost and risk
- Value hypothesis and counterfactual
Where will released capacity go?
- Planned destination: work, demand, owner and acceptance
- Realised destination: observed use, owner, outcome and date
- Variance: none / minor / material; reason and decision effect
Management test
- Case: what result matters, for whom, and against which baseline?
- Net change: what became faster after review, rework and recovery?
- Usable capacity: can it be scheduled at the constraint?
- Destination: where did it go, who owns it and what differed from plan?
- Enablement: skills, authority, access, tools and incentives?
- Accepted outcome: what occurred at required quality and who accepted it?
- Attributed value: what remains after costs, alternatives and counterfactual?
Classification
- Benefit: cash / avoided cost / throughput or growth / quality or customer / risk or resilience / workforce or learning / strategic option
- Recipient: enterprise / customer / employee / partner or ecosystem / society
- Evidence: estimated / self-reported / scenario / open / disputed / stale / externally reviewed
- Risks and reversibility
Decision
- Pilot / redesign / expand / stop — conditions and accountable approval
- Action, owner and deadline
- Approver and acceptance evidence
Do not value the hour. Value the accepted outcome.
Each stage needs its own evidence: claimed, measured, checked, corroborated or tested against a counterfactual. Accepting an outcome does not strengthen its evidence.
© 2026 ETHORITY & STEN.AI LLC
A REAL SOURCE ARTIFACT
Keep the outcome, evidence and decision together.
The original one-page record asks for the workflow, owner, outcome acceptor, baseline, planned and realised capacity destination, evidence and next decision. It is a blank method instrument, not a completed customer record.
Download worksheet · PDF ↗Worksheet adapted for readable web and print use from the source method. Request the applicable rights for redistribution or adaptation.
Research, application and commercial interests
ETHORITY offers advisory and implementation services based on this method. That interest is disclosed; it does not alter an evidence code or justify a favourable result. Independent review is identified separately from ETHORITY’s own analysis.